The study focused on the conventional yet comprehensive use of forecasting in sales prediction.  Businesses use forecasting to ascertain the allocation of their budgets or schedule for foreseen expenditures for a future time frame. In this study, forecasting is placed in the context of SG Trading which is a Filipino-owned company operating in Quezon City, Metro Manila.  Methodically, the study makes use of both literature and actual experiment. The review was employed to check the prevailing existing significant literature to summarize the results. Similarly, the experiment was considered because it is the most satisfactory research method when exploring quantitative data (Brownlee, 2020). Its main goal was to gauge the performance of the classical statistic forecasting model, namely Auto-Regressive Integrated Moving Average, and Autoregressive and moving average, machine learning models, Support Vector Machine Regressor, Random Forest Regressor, and Gradient Boosting Regressor on the sales data obtained from the sales daily report of SG Trading. The results from the experiment revealed that XGBoost is the best suitable algorithm to forecast the LPG daily sales of SG Trading with the least error. In view of such findings, it is recommended to adopt and utilize the said forecasting model so as to have an effective planning of the LPG supply. Part of the limitations was the lack of variables; hence, the researchers have also suggested enhancing the sales data. In a nutshell, it is strongly enjoined to utilize the findings produced by various forecasting models for the improvement and sustainability of the business.